Large-Scale Data Reporting of Paediatric Morbidity and Mortality In
Total Page：16
File Type：pdf, Size：1020Kb
Global child health Arch Dis Child: first published as 10.1136/archdischild-2015-309353 on 21 October 2015. Downloaded from Large-scale data reporting of paediatric morbidity and mortality in developing countries: it can be done Trevor Duke,1,2 Edilson Yano,3 Adrian Hutchinson,4 Ilomo Hwaihwanje,5 Jimmy Aipit,6 Mathias Tovilu,7 Tarcisius Uluk,8 Theresia Rongap,9 Beryl Vetuna,10 William Lagani,11 James Amini,12 on behalf of the Paediatric Society of Papua New Guinea ▸ Additional material is ABSTRACT are that almost no standardised data are available published online only. To view Although the WHO recommends all countries use from non-tertiary health facilities in low/ please visit the journal online fi (http://dx.doi.org/10.1136/ International Classi cation of Diseases (ICD)-10 coding middle-income countries, and the data that exist archdischild-2015-309353). for reporting health data, accurate health facility data comes from tertiary facilities that have diagnostic- are rarely available in developing or low and middle based funding models or insurance schemes. At the For numbered affiliations see end of article. income countries. Compliance with ICD-10 is extremely same time as recommending this complex coding resource intensive, and the lack of real data seriously system, at the clinical level WHO recommends Correspondence to undermines evidence-based approaches to improving simpler syndromic diagnoses for common childhood Professor Trevor Duke, Centre quality of care and to clinical and public health conditions, based on the Integrated Management of for International Child Health, programme management. We developed a simple tool Childhood Illness (IMCI), Hospital Care for University of Melbourne, MCRI, 34 Royal Children’s Hospital, for the collection of accurate admission and outcome Children and other disease-specific guidelines. Flemington Road, Parkville, data and implemented it in 16 provincial hospitals in We sought to develop a tool for recording and VIC 3052, Australia; Papua New Guinea over 6 years. The programme was monitoring admissions of children to hospitals that [email protected] low cost and easy to use by ward clerks and nurses. could be implemented on a large scale in settings Received 17 July 2015 Over 6 years, it gathered data on the causes of 96 998 with very limited resources. The programme Revised 7 September 2015 admissions of children and 7128 deaths. National requires only basic patient demographic informa- Accepted 10 September 2015 reports on child morbidity and mortality were produced tion and diagnoses that are routinely recorded in Published Online First each year summarising the incidence and mortality rates admission/discharge record books kept in most chil- 21 October 2015 for 21 common conditions of children and newborns, dren’s wards. The diagnoses had to be consistent and the lessons learned for policy and practice. These with those used by the WHO in clinical guidelines, data informed the National Policy and Plan for Child and consistent with those in ICD-10 coding Health, triggered the implementation of a process of systems, but less numerous, not reliant on labora- clinical quality improvement and other interventions to tory tests that were not available, and the diagnoses reduce mortality in the neediest areas, focusing on had to be easy to identify and record by ward diseases with the highest burdens. It is possible to clerks or nurses with minimal training. collect large-scale data on paediatric morbidity and In 2007–2008 we designed the Paediatric http://adc.bmj.com/ mortality, to be used locally by health workers who Hospital Reporting (PHR) programme and progres- gather it, and nationally for improving policy and sively implemented it on a national scale. We practice, even in very resource-limited settings where report the development, implementation, and out- ICD-10 coding systems such as those that exist in some comes of the PHR over 6 years in Papua New high-income countries are not feasible or affordable. Guinea (PNG). on September 27, 2021 by guest. Protected copyright. METHOD INTRODUCTION Design characteristics and processes Since 1967, the WHO has recommended that all The programme was developed in PNG and member states use the International Classification Australia. In 2008 a meeting of all PNG paediatri- of Diseases (ICD) coding. However, by 2005 only cians, Ministry of Health officials and academics 23 countries were considered to have high-quality proposed the diagnoses to be included in the initial death registration data (quality criteria based on version. These included all conditions in the WHO Open Access Scan to access more timeliness, completeness, coverage and the sparing IMCI and Hospital Care for Children guidelines, free content use of codes for ill-defined causes).1 These were and the PNG Standard Treatment Manual for wealthy countries using ICD-10 coding for cost Common Illnesses.34The diagnoses were standar- reimbursement health financing systems based on dised, consistent with WHO definitions and diagnoses. ICD-10, although numerical coding was not used. Developing or low and middle income countries The programme was made using FileMaker Pro have not kept pace with diagnostic precision required (http://www.filemaker.com/) at a development cost for optimal classification using ICD-10, and the of less than US$10 000 spread over 6 years. To cite: Duke T, Yano E, resource requirements for ICD-10 conformity are Because the manual calculation of basic statistics Hutchinson A, et al. Arch considerable.2 It is particularly beyond the capacities in health reports is often incorrect, the programme Dis Child 2016;101: of poorly resourced provincial and district hospitals needed to automatically calculate the number of – 392 397. in low/middle-income countries. The consequences admissions for common diagnoses, the number of 392 Duke T, et al. Arch Dis Child 2016;101:392–397. doi:10.1136/archdischild-2015-309353 Global child health Arch Dis Child: first published as 10.1136/archdischild-2015-309353 on 21 October 2015. Downloaded from deaths and the overall, age- and disease-specific case fatality Table 1 Total admissions and outcomes for each hospital 2009– rates (CFRs). Furthermore, most health information systems 2014 record only one diagnosis; this leads to a significant underesti- mation of common comorbidities, such as malnutrition, HIV Overall case and anaemia. The PHR therefore needed to record comorbid- No. of years Total Total fatality rate Hospital reporting admissions deaths (%) ities; without any double-counting of patient numbers, the auto- mated calculations help to understand the contribution of these Alotau 2 2492 49 1.97 comorbidities to morbidity and mortality. The programme also Angau 4 8672 1016 11.72 needed to include simple illness severity metrics so that differ- Buka 6 3167 261 8.24 ences in CFRs over time and differences between health facil- Daru ities may be better understood. These included severe Goroka 6 16 876 919 5.45 pneumonia, which has a standardised definition that is well Kavieng 3 1084 63 5.81 4 understood by health workers, and very low birth weight, an Kimbe 5 5242 514 9.81 internationally accepted classification which uses an objective Kerema metric of weight (1000–1499 g). Kundiawa 2 4695 342 7.28 Manus 3 988 16 1.62 Implementation and system requirements Mendi 2 4405 235 5.33 Some hospitals had a computer in their wards; others required Modilon 6 8063 794 9.85 purchasing of a computer for use of the programme. The pro- Mt Hagen 4 15 839 1129 7.13 gramme is designed for Windows-based operating systems and is Nonga 4 3088 213 6.90 installed via self-contained executable file either downloaded or Popendetta 2 2810 216 7.69 distributed via universal serial bus or compact disc. Port Moresby 3 12 976 911 7.02 One day of hands-on training was provided for ward clerks, Vanimo 3 2290 96 4.19 nurses and doctors in the 16 hospitals to enable them to use the Wabag 5 3265 276 8.45 programme. A printed discharge form was used as part of the Wewak 1 1046 50 4.78 routine medical record (see online supplementary appendix 1) Total 61 96 998 7128 7.35 and completed by the doctor or nurse discharging the patient. The data recorded were name, date of birth, address, weight, discharge diagnoses and outcome. The data form enabled a ward clerk or health worker to enter patient data after the and per common disease are illustrated in figures 1 and 2, child’s discharge from the facility. Other data such as complete- respectively. Not all hospitals were able to report each year. The ness of vaccination was added in the latest versions of the pro- numbers of hospitals that were able to report each year were 7, gramme. The data entry programme needed to be intuitive, 11, 11, 10, 10 and 12 in 2009 to 2014, respectively. This with minimal number of screens to navigate, and data entry amounted to a total of 61 hospital-years of complete reporting. needed to take less than 1 min per patient. Reasons that hospitals were unable to report fully in any given The data were collected and summarised in individual hospi- year were often computer problems (such as a virus that shut tals, where outcome data were automatically calculated on a down a computer), or staffing problems (ward clerk leaving and summary sheet for any given time period (example in online a gap between appointments). http://adc.bmj.com/ supplementary appendix 2). This enabled the data to be used at The PHR programme was updated several times over the a local level to monitor activity and disease patterns, for audit- 6 years; faults were corrected, and new diagnoses were added ing and to plan local interventions. Each year the summary data based on consensus among the Paediatric Society and the from each hospital were collated at the National Department of Health Department.